Overcoming Theory: Designing Brainwave Authentication for the Real World

Overcoming Theory: Designing Brainwave Authentication for the Real World

연구 분야: Analysis

학회: EuroUSEC '23: Proceedings of the 2023 European Symposium on Usable Security

Recent advancements in consumer-grade Brain-Computer Interfaces (BCIs) have opened up new possibilities for the exploration and deployment of brainwave-based user authentication. However, research on real-world usability and security is still lacking. Additionally, existing prototype implementations are not widely accessible. To address these limitations and facilitate future prototyping, we developed NeuroPack, a Python library tailored toward the prototyping of brainwave-based authentication systems. Contained within the library is KeyWave, an easy-to-use and customizable example implementation allowing for continuous authentication using Electroencephalography (EEG) signals. We integrated KeyWave into a password manager and conducted a preliminary user study with nine participants to assess its usability and security, while also investigating the general acceptance of users toward this new technology. The results of the study reveal high usability scores on the System Usability Scale (85.28) but also highlight possible concerns regarding user willingness to use a separate recording device for authentication. Our results further indicate, that existing machine learning models perform well with real-world data, showing promising results with an Equal Error Rate of 7.95%.